THE INFLUENCE OF GEOLOGIC SAMPLE SURFACES ON TIR SPECTROSCOPY
Evelien Rost is a PhD student in the department of Applied Earth Sciences. (Co)Promotors are prof.dr. F.D. van der Meer, dr. C.A. Hecker and dr. H.M.A. van der Werff from the Faculty Geo-information and Earth Observation (ITC).
Thermal infrared spectroscopy is a robust method to quickly and reliably measure and analyze large geological datasets. However, spectral measurements of geologic materials (both solid surfaces and powders) can be significantly influenced by various sample characteristics such as surface roughness, particle size, porosity, fine surface particles and sample and/or mineral orientation. The relationship between these sample characteristics and the spectral changes is still poorly understood.
This thesis combines empirical-based laboratory research, novel non-linear spectral modelling and an interlaboratory experiment to gain insights in the processes that affect the spectral signature. The objective of this thesis is to obtain an understanding of how rock sample surface preparation influences the spectral shape and spectral contrast and identify if comparable spectral changes and trends can be observed for various spectroscopy devices, including three different TIR spectroscopic techniques (directional hemispherical reflectance, bi-directional reflectance and emissivity).
Chapter 2 explores the effect of sample surface preparation by comparing thermal infrared spectra of three sample surface preparation methods (polish, saw and split). The influence of the surface roughness is two-fold, one effect is a changing spectral shape, and the other is a varying spectral contrast. The first effect is consistent for each applied spectrometry method, while the second effect is more variable depending on the measuring technique.
To improve understanding of the physical processes that influence the observed spectral variation, chapter 3 presents a non-linear spectral model that combines rock surface reflection with transmission through clinging fines (grain size <10 µm) that coat the surface. The novelty of this approach is that it does not model individual particles, but instead compiles different combinations of optically thin particle transmission and reflectance of optically thick mineral reflectance spectra until an optimal fit with the actual measured rock spectra is achieved. The model manages to reproduce the spectral amplitude and shape of the strong bands of the quartz reststrahlen doublet of three of the four sandstone spectra, which proved challenging for previous models.
To identify the impact of the sample surface preparation on the comparability of thermal infrared spectral datasets, an interlaboratory experiment is presented in chapter 4 that compares spectra of seven thermal infrared spectroscopy laboratories. The results show that for samples with simple mineralogy, sample surface preparation is not a large factor and the spectra are mainly impacted by measuring technique and differences between individual spectrometer properties. While for samples with a more complex mineralogy the sample surface preparation has a much bigger impact, although the statistical analyses still show the impact of the measuring technique as well. Furthermore, the results of both the spectral measurements and the spectral statistics show two distinct groups of spectrometer devices, indicating that direct comparison of thermal infrared spectra of different spectroscopic techniques is still challenging.
The research presented in this thesis identifies some of the key components needed to improve usability of the TIR wavelength range for geologic applications. It provides insights on processes that impact spectral signatures and identifies opportunities for spectral correction by formulating a best-practice-path for rock sample preparation that ensures consistent and comparable TIR spectral signatures.
